Skip to main content
Version: Next

Troubleshooting / FAQ

The execution status does not progress​

Execution StatusExecution StatusExecution Status

The import was prepared and the records are queued, but no worker processes them. Set up either command-based or Symfony Messenger-based queue processing, as described in Installation.

Check that the worker covers the execution type the configuration uses. A configuration set to Parallel is not processed by datahub:data-importer:process-queue-sequential, and the other way around. See Import Execution Details.

"ERROR: The command is already running." when processing the queue​

datahub:data-importer:process-queue-sequential takes a named lock for the duration of its run, so only one instance processes the sequential queue at a time. The message means the lock is held. That is the expected outcome while another invocation is legitimately still running, for example a long import overlapping with the next cron tick, and it is not by itself evidence of a problem.

Confirm the lock is stale before touching it. Check the host for a running process (ps aux | grep process-queue-sequential) on every machine that runs the command, and check whether the import is still making progress in the Execution tab. Releasing a lock that a live worker holds lets a second processor run concurrently, which is exactly what the lock prevents.

If no process is running, the lock is stale and expires on its own after the command's 24 hour TTL. To clear it sooner, delete it from the lock store configured for the installation (framework.lock in the Symfony configuration). With the default database store that is the lock_keys table, where the key column holds a hash of the command name. Other stores (Redis, filesystem, Zookeeper) keep the entry elsewhere, so check the configuration rather than assuming a table exists. Then run the command again.

A scheduled import never starts​

Cron and one-time schedules are evaluated by datahub:data-importer:execute-cron. If that command is not run regularly, no scheduled import starts. See Installation.

A one-time job is also skipped when the configuration was saved after the scheduled time, and it never runs a second time.

An import creates duplicates instead of updating objects​

The element loading strategy did not find the existing object. Check in Resolver Settings:

  • The Data Source Index points at the field that actually carries the identifier.
  • With the Attribute strategy, the existing objects are published, or Include unpublished objects is enabled. Unpublished objects are not matched by default.

An import reports that the preview file is invalid​

The preview file does not parse with the configured file format, usually after the format was changed. Re-upload or re-copy the preview data. See Import Preview.

Pushing data returns an error while the queue is filled​

An import only prepares when the queue of that configuration is empty. Either let the running import finish, or enable Ignore Not Empty Queue on the Push data source.